 |
 |
Archives of the TeradataForum
Message Posted: Tue, 23 Jan 2002 @ 01:46:16 GMT
| Subj: | | Re: Pondering |
| |
| From: | | Neil Raden |
Actually, to set the record straight, there are more than 2 vendors out there. IBM is starting to pick up a fair share of the very large
data warehouses in the open systems area (and there are already quite a few very large ones in DB2 on the mainframe), and I've come across a
few lately in, believe it or not, Sybase IQ. I haven't run into much Informix or Red Brick lately (but I have clients how are still using
both and are quite happy), but that's all IBM now. Tandem's NonStop SQL/MP is gone too, and for those of you who worked with it, you
probably miss it as much as I do, it was a great product.
If you think about this logically, though, it doesn't really matter how large (or small) your data warehouse is. After all, all that data
just sits on disk drives. What is important is how much of the data is processed by the database server. That's where your requirements
help you understand what kind of product is needed. So, theoretically, you could use SQL 2000 with a 4TB warehouse if most of the data was
inactive; you might need Teradata for 500Gb warehouse is every query exercised a good chunk of the data.
But I've been watching all of these replies and I think we've all overlooked something. Sure, if the cost of ownership is less, that's an
important issue, but aren't data warehouses built for people, not for DBA's? I'd like to hear some discussion about why Teradata is better
for the users and, especially, for being integral to the effort to change the way people work with information. If you want to be cynical
about it, you could say that SQL 2000 is the best database because people can run their queries from within their speadsheets. How about
some serious thought about this?
-NR
| |
